Skip to content
Portland Civic Lab
Support
← Back to the guide
The rules behind the guide

Fair to candidates.
Accountable to the evidence.

Nonpartisan means using the same standards for everyone. It does not mean treating a documented fact and an unsupported assertion as equally credible.

Research reviewed September 18, 2026. This edition contains AI-assisted research and drafting; it has not completed a separate human editorial review.

What we are trying to explain

For each candidate: what they want government to do, how they propose to do it, what their public record can establish, and what remains unanswered. We describe tradeoffs and institutional limits. We do not endorse candidates, recommend rankings, assign ideological scores, or infer private motives.

The same questions for every candidate

  1. What specific policy or administrative changes do they propose?
  2. Does this office have the power to make those changes?
  3. What would they cost, who would pay, and what would be reduced?
  4. What evidence supports their factual claims and claims of success?
  5. What result and deadline would allow the public to evaluate the promise?

The same questions do not require identical conclusions or word counts. A concrete policy warrants more explanation than a slogan. A documented falsehood warrants a correction regardless of who said it.

Three kinds of writing, visibly separated

Candidate statements

Campaign websites, interviews, questionnaires and candidate-authored voters’ pamphlet entries establish stated positions. Hosting a statement on a government website does not make its claims verified. We paraphrase with attribution and do not convert a promised result into an established fact.

Public records

We check the actual bill, roll call, budget, audit or official decision. We distinguish introduction from passage, a procedural motion from a final vote, and an appropriation from an achieved outcome. A single vote is not a complete political identity. In this edition, all six Portland council incumbents are compared on the same 73 decisions across 29 issues. Every subject in the Council site’s current 24-topic index is covered, alongside additional major votes on labor, children’s services, policing, elections and City governance. Each decision includes the same six incumbents, with amendments, absences and committee membership clearly distinguished. This is a comparison of consequential choices, not a score.

We follow competing budget proposals and Zenith’s successive stages, alongside the amendments that shaped the Moda negotiations. Personal explanations are attributed to the councilor or reporting that identifies the speaker. Earlier statements keep their dates. A documented action is labeled as an action when we have not established an individual explanation; we do not turn a vote into an inferred motive. The four featured votes on each race page cover the district’s sitting councilors; the candidate list and issue lines include every candidate.

How we checked Council topic coverage

We reviewed the Council site’s 24 topic dossiers and scanned 339 unique final-action entries in its vote index. We then checked selected measures against the City’s official legislation and roll calls, rather than treating topic summaries or index dates as the authoritative record. The downloadable evidence includes a map from every dossier to its issue comparison here.

Coverage means that every indexed subject has a matched account; it does not mean every motion, hearing or public statement appears in this guide. Routine transactions are not all reproduced. For pending matters, we describe the latest verified stage: the firearms proposal was tabled in committee, natural-psychedelics legislation awaited full Council consideration, and Zenith reconsideration remained scheduled for September 23. Members outside a committee are marked “Not on committee,” never “Absent” or “No.”

Editorial interpretation

We explain the practical meaning of the supported position: for example, a temporary moratorium leaves a different decision open than a permanent prohibition. Every interpretation must be traceable to the cited evidence. We do not turn party affiliation, endorsements, donors, identity or silence into a position the candidate has not expressed.

Selection, fairness and context

The comparison’s values labels describe priorities supported by the cited materials, not personal virtue, hidden motives or fixed ideological identities. Readers can compare the same issue across candidates without a score or recommendation. An unfilled issue means the reviewed evidence does not establish a position; it does not mean neutrality, opposition or lack of interest. More detailed campaign writing is not itself evidence of stronger performance.

  • Include every named candidate on the checked roster for a covered race, including smaller-party candidates and candidates with little public material.
  • Display candidates alphabetically by the name shown, with the same visual treatment. No candidate is preselected for comparison.
  • Give incumbents an independently checked record where available. Do not penalize challengers for lacking roll-call votes; look for relevant professional, civic and public statements instead.
  • Use complete source context. Before publishing quotations from audio or automatic transcripts, verify the recording, speaker, date and timestamp. This edition does not quote unchecked transcripts.
  • Distinguish current platforms from historical statements. A change of position needs dated evidence on both sides and, where available, the candidate’s explanation.
  • Do not publish a serious allegation as fact on an opponent’s say-so. Establish the underlying evidence and seek the subject’s response before adding disputed allegations.
  • Endorsements and contributions may illuminate alliances, but do not establish motive, control or agreement with every position held by a supporter.

Photographs and equal presentation

Portland Council portraits use candidate-submitted voters’ pamphlet photographs wherever available, with additional images from identified campaign, civic-event or news sources. Each brief links to its photo source. Images share the same frame and receive no synthetic retouching. A placeholder means a suitable portrait has not yet been sourced; it is not a judgment about the candidate. Published photographs remain the work of their respective owners.

Portland Council: what is ready

Both districts have complete checked rosters: 21 candidates in District 3 and 12 in District 4. Thirty-one have substantive current policy briefs. Darren McCormick has a limited current filing statement; Pham’s current issue pages are now reviewed separately from his historical 2024 answers; John J Goldsmith has verified current background but needs additional platform research. Thirty-one Council portraits are sourced; photographs for McCormick and Goldsmith remain outstanding.

This supports an openly labeled initial research edition. It does not establish a complete investigation of each candidate’s public record. The six incumbents have seventeen common roll-call checks; additional topics such as transportation funding, Street Response, the oversight system’s design, water infrastructure, labor and governance still need matched candidate accounts. The race pages link to the Council dossiers for this wider context. Independent editorial review remains incomplete.

Coverage, honestly stated

48races in this edition
146/153candidates with substantive briefs
14profiles with a separate record check

The public edition covers Portland City Council Districts 3 and 4 only. Both elect three councilors in November 2026. Research for Multnomah, Washington and Clackamas counties and major state and federal races is in development and has not been published here.

This is a working research edition, not a claim to have reviewed everything every candidate has said. Ballot measures and other offices are outside this first release. The coverage count describes published entries, not all races on your ballot. Pamphlets are not complete candidate rosters: candidates can decline to submit a statement.

Known gaps in this edition

Across the remaining profiles, campaign finance, complete roll-call histories, past-position changes, independent outcome verification and full debate/interview review are not yet complete. On September 19, 2026, we emailed all 33 candidates in the published District 3 and 4 races with individual research questions and an invitation to correct or clarify their profiles. Sending does not establish delivery or a response. An open question is our research question, not a claim that a candidate refused to answer it.

Beyond the Council: the same format for every office

Every race on a Portland ballot, and every race we have researched in the metro area and the state, uses the same page: the same four issues, the same chips, ladders, verbatim openings and campaign contacts, alphabetical, with gaps shown as gaps. Two things differ by office. Only Portland City Council seats carry the recorded-votes panels, because that is the record our research checks vote by vote; other incumbents’ recorded actions appear on their briefs under “Public record” where our sources establish them. And the extra comparison topics are Council choices, so they apply to Council races only for now.

The state voters’ pamphlet is published on September 29, 2026. Until then, governor, congressional and legislative candidates have no pamphlet statement, so their openings come from their campaign sites under the same rule, and their portraits are campaign photos credited to the site. We will reconcile every statewide and legislative brief against the pamphlet when it appears.

What’s at stake, and the size of each seat

Each race page carries “What’s at stake”: the office’s biggest current problems and pending decisions as facts with a number or a date and an official or reported source, the same block for every candidate. Beneath it, each of the office’s own live choices is a board: the question, then every candidate A–Z with a recorded vote, an explicit statement or the gap. A board row is filled only by an explicit statement or, for a sitting official, a recorded action cited to the record, and is labeled which; a statement that speaks to the topic but not the exact choice is shown and labeled as such. The four columns of the grid keep the same four issues for every race, worded for the office (the county’s “Crisis care, jail and deflection” is the Council’s “Camps, crime and who responds”). On the hub, each body’s bar is its most recent adopted budget on a logarithmic scale, with the population it serves and who decides, each figure linked to the budget document or the Portland State University population estimate it came from.

Reaching a campaign

Each brief lists how to reach the campaign: the website, email, phone, contact form and social profiles the candidate published in their voters’ pamphlet statement or on their campaign site, or the campaign email on their own City filing or announcement. We publish nothing a candidate did not publish for the campaign, never a home address, and nothing from third-party directories. When a candidate has published no channel, the brief says what we checked.

How the race page works

Each race is one grid: every candidate on the checked roster, alphabetical by displayed name, against the same four issues (rent and homes; camps, crime and who responds; your bills and taxes; streets, buses and air). Each cell is a chip of two to four words: our short reading of that candidate’s documented position. Tapping a chip shows the sentence it was shortened from, the full position with any qualifier, and the source. A dash means we have not established a position from the sources we reviewed; it is a research gap, not a position, and each column header says how many candidates have a documented position. Columns are colored by topic, never by side.

Beneath each chip is the same three-rung ladder for every candidate: what they say they would do, how they say they would do it, and how they say we would know it worked. Each rung has its own source. A rung we could not find in a candidate’s pamphlet, website, questionnaires or reporting is shown as a gap, and where we have emailed the candidate to ask, the date of that request is shown beside it. We never fill a rung for a candidate.

Beneath the grid, each concrete choice the Council has faced (the Moda Center terms, new taxes or fees, police staffing, camp removals, data centers, Portland Street Response, the street repair fee, water rates) is a board of its own: the question, then every candidate A–Z. The list and its order are editorial and the same for every race. Where a topic matches a recorded Council decision, sitting councilors’ votes appear as pills. Every other row is either an explicit, sourced statement about that exact choice, a statement that speaks to the topic but leaves the exact choice unsaid (shown and labeled as partial), or a gap; we never infer a stance from party, silence or a broad goal, and each board says how many candidates are on record and, in a small field, who is not.

Chips and short lines are written by us, shortened from the full position, and each names the field it was shortened from. They may omit; they never add. The full position, its source and any qualifier are one tap away on the candidate’s card and brief. Candidate statements carry a quotation glyph; the source chip is on the card and brief. Recorded votes carry a ballot glyph. Our interpretation is always labeled “Our reading.”

Four featured votes appear on each race page: the most reported and most divided decisions among that district’s sitting councilors, phrased as the question Council decided, with each councilor’s recorded vote (Yes, No, Absent or Not on committee) and the decision’s limits. Beneath each, challengers who have made an explicit statement about that exact choice are listed alphabetically with a short attributed paraphrase and its source; candidates without such a statement are counted, not characterized. The full record of every topic remains on the Votes page. Challengers have no Council votes; that is not a judgment.

“My ballot” is the reader’s own list. You choose which names to save and in what order, and you may add a private note. The site never fills, sorts, scores or shares that list; it is stored only in this browser tab and never appears in a link, a screenshot label or our analytics. A shared link can open a view of named candidates, but it never writes to your list. There is no quiz, match score or recommendation. The earlier alignment quiz remains retired and marked as such in the evidence export.

Page views, including candidate brief pages, are recorded by our site analytics; we configure our analytics not to receive URL fragments, and your saved list and issue choice are not recorded. Every original profile, decision and issue analysis remains available on the brief, Votes and print pages and in the evidence export, which now includes the short-line layer with its version so reviewers can compare it with the underlying research. Independent editorial review of the short lines and first-time-voter usability testing remain outstanding. The issue views went live on September 19, 2026 before that review, and each race page says so until a reviewer signs the lines.

Sources and reproducibility

Each profile links to its evidence, with document dates or a clearly identified review date. PDF links use the PDF page number, which can differ from the printed page label. Retrieved sources are retained as research working files. The manifest hashes original PDF and spreadsheet downloads and normalized extracted text for web pages; it identifies which representation each checksum covers. Policy briefs are dated editorial snapshots, not a live feed.

Read the research and corrections log · Download the briefs and evidence links (JSON)

Corrections and candidate responses

Use the Lab’s contact form with the race, candidate, exact sentence, proposed correction and public source. Candidates and readers use the same evidence standard. A campaign may correct a factual error or add a sourced response; it does not control the guide’s analysis.

Material corrections should identify what changed, why, the evidence and the date. The initial research edition has no published correction entries. Subsequent changes belong in the public research log rather than silently rewriting the record.

Independence

This guide is produced by Portland Civic Lab, an independent organization unaffiliated with election authorities. It contains no paid candidate placements or endorsements. Read the Lab’s independence and funding policy. Any relevant financial or personal conflict discovered during editorial review must be disclosed before a profile is described as fully reviewed.